Magnetic resonance force microscopy (MRFM) is a scanned-probe imaging technique interrogating quantum spin relaxation in a magnetic field gradient. Whereas classical nuclear magnetic resonance (NMR) relies on ensemble spin resonance detected by an inductively-coupled antenna, force-detected MRFM enables the Larmor resonance condition to be satisfied in a much smaller volume, thereby enabling the back-action of a single spin on a magnet-tipped cantilever to modulate the cantilever resonance. This enables single-spin detection, and stands as a path to non-destructive, three-dimensional mapping of nanostructures, complex synthetic molecules, and biopolymers.
Maximal signal-to-noise ratio in an experiment like this requires a particular cantilever: soft, yet spectrally pure, cantilevers tipped with the smallest possible ferromagnets increase SNR while controlling losses coupled into dielectric fluctuations in the sample. Moreover, high-frequency operation – ideally in the megahertz – provides faster averaging and potentially new imaging modalities, with direct coupling of the cantilever resonance to the nuclear spin precession frequency.
As an sophomore, my work consisted of designing, fabricating and testing rf-MEMS cantilevers to fulfill this panoply of – at times orthogonal! – design goals. To this end, my first technical work was assisting in the fabrication of singly-clamped magnet-tipped cantilevers at the CNF. As an undergrad, I assisted at every point in the process, using a combination of well-controlled electron beam lithography and anisotropic wet etching to fabricate sharpened bulk-silicon tips, capped with a single cobalt magnet. I came to regularly write feature sizes below 100 using the state of the art JEOL JBX-9300FS tool available at the CNF, made electrical contact to them using an i-line stepper photolithography process, and etched the resultant structures with plasma and wet etch chemistries.
These abilities opened an opportunity to head a project that further leveraged the resources at the CNF. In order to permit experimental modalities that couple the cantilever resonance to the spin precession frequency, the Marohn group was interested in the development of megahertz mechanical cantilevers. This would be an increased challenge, since packaging considerations preclude the ability to use the typical fiber-optic interferometer in the cryostat – a different mechanism would have to be implemented to observe the cantilever displacement, described below.
To begin this effort, I developed a fabrication protocol for silicon nitride, gold-coated megahertz beams. As an undergrad, I catalogued, proofed and implemented fabrication recipes for the relevant metal and dielectric depositions, the stepper and electron beam lithography, the silicon nitride etch and the HF release of the beams. I developed an electron beam lithography recipe to make the 120 wide beams, and made contact to them with an aligned i-line photolithography step. I developed a liftoff precess to metallized the beams, and etched them with an RIE etch. Finally, I used critical point drying to release the beams.
To characterize these devices, I used a Lorentz-force magnetomotive drive to observe cantilever motion in vacuum under cryogenic conditions. In this experiment, a magnetic field normal to the plane of the cantilever resonance couples to free charge carriers in the metallic layer. These carriers, drifting under time-dependent applied potential, apply force to the beam and excite detectable mechanical resonance, observable as a time-dependent change in the impedance.
I implemented an actuation and detection methodology using radio-frequency lock-in amplifiers and network analyzers, engineered the impedance-controlled coupling from the cryogenically cooled device to the electronics, and characterized the cantilevers’ response to increased magnetic field and drive frequency. I measured the beam’s response at cryogenic temperatures as a function of applied magnetic field, and observed an appropriate quadratic relation.
While doubly-clamped beams are readily observable at resonance using the magnetomotive methodology, our architecture for the MRFM experiment used a singly-clamped cantilever. In order to provide better packaging and to minimize stray power to the device – and avoid heating above the 4.2 K operating temperature – we implemented a detection scheme for the cantilever position using tunnel current fluctuation sensing. With this approach, we monitored the time-dependent tunnel current between a fixed electrode and a patterned electrode on the cantilever itself, while driving the cantilever with the Lorentz force experiment described above.
The most memorable challenge we overcame was the reliable fabrication of a high aspect-ratio structure in metal using electron beam lithography, with dimensions below 15 nanometers. This required an exquisite understanding of the limitations of electron beam lithography, which I by then possessed. Even more challenging was the etch and release of the resultant structure, while preserving the small dimensions. However, I solved these problems each in turn, before developing a technique to electron-migrate the gold overlayer to form the tunnel gap.
We used the doubly-clamped cantilever measurements as a control, and observed substantial fidelity between the two metrology modes. This work was presented at the at the 2006 Kavli Institute at Cornell Summer School in MRFM.
As an undergrad, I spent over 2000 user hours at the CNF, and became proficient at the rapid fabrication of nanoscale MEMS devices. I learned about pitfalls of various fabrication processes, and put this knowledge into practice by fabricating rf-MEMS cantilevers. I was the youngest user of the JEOL 9300 electron beam lithography tool, and regularly operated the tool at the limits of the feature size for the geometry we had in mind.
I also became proficent at testing devices. I designed and built a Lorentz-force based MEMS beam testing setup, becoming familiar with vacuum hardware and RF electronics. I presented the results I achieved with this experiment at conferences.
Most importantly, this early work at the CNF confirmed that research was indeed what I wanted to do. The realities of hunting down hypothesis, planning experiments, wrestling hardware and forming conclusions were made concrete; I was excited by the challenges they provided.